[ricecrispi]

Bill mention standards in testing. How can we standardize something like a TIM that is variable and outside parameters of control? Why should testers have to account for it in testing then? Yeah TIMs exist and they affect performance but we are testing WBs and not TIMs. TIMs vary too much and influences results. We want accurate results of WB performance in ideal situations, not one where the TIM is much of a factor in performance as the WB. We can vary the die size and I think that good enough. You guys can add the TIM caps to the dies but using realife CPU for testing is purpose is flawed. CPUs vary and so do TIMs so they are fine for comparative results but not pure performance.

I understand Bill's position. The TTV is a more realistic test setup than a die sim so it's would seem to get results more reflective of real life results. That doesn't make it more accurate or reliable which is Cathar's standpoint I believe.

Swiftech obvious knew about TIM variablity and tried to address it with the flex base idea. Why else would they put that statement down. It could be BS, which GABE seems to be marketing, or maybe they accurately addressed the situation with TIM variablilty but the results dont show in a die sim. It still doesn't mean the block performed better than the Storm because it address a TIM issue better.

It's like racing on a track with potholes or a bad surface. Just because a car can handle the potholes better doesn't mean it's a faster car. It's just a a car that run well with tracks with potholes.